Aileen M Aldrich^1*, Elizabeth R Lyden², Jenny Edouard³, Jacqueline Gautier³ and Shirley F Delair^1
*Correspondence: Aileen M Aldrich, Division of Pediatric Infectious Diseases, College of Medicine, University of Nebraska Medical Center, Omaha, Nebraska, USA, Tel: +14029554005 extn. 5, Fax: +14029553849, Email:

Keywords

Global health; Haiti; Healthcare quality indicators; HIV; Pediatrics

Introduction

Pediatric HIV burden in Haiti

Haiti is home to the largest pediatric human immunodeficiency virus (HIV) population in the Americas. As of 2017, there are roughly 150,000 people living with HIV in Haiti 7000 of which are children under the age of 15 [1]. In 2004, the United States President’s Emergency Plan for AIDS Relief (PEPFAR) began providing free antiretroviral therapy (ART). Early infant diagnosis (EID) for HIV has also been expanded since the introduction of PEPFAR; the country has gone from 9 sites where EID testing was performed in 2009 to 106 sites in 2014 [2]. Based on the latest pediatric HIV guidelines for Haiti, all inpatient and outpatient clinics should provide HIV testing as well as testing for all malnourished children and children receiving immunizations [2,3]. In 2017, of all infants <12 months who were screened for HIV, the total transmission rate was 5.7%, which was a slight increase from the prior year [1]. According to PEPFAR’s 2018 country report, about 3530 pediatric patients aged less than 15 years were receiving ART, however, this number only represented about 50% of the infected children [4]. There is a need for continued expansion of ART coverage to all infected children; this process, however, should be accompanied by comprehensive quality of care measures, to ensure optimal care in the context of limited resources.

The proposed model

An available quality of care tool that has been used in developed countries includes the HIVQUAL model [5-7]. This model was created by the New York State Department of Health Acquired Immune Deficiency Syndrome (AIDS) Institute as the first set of framework for assessing adult HIV quality of care measures [8]. In 2003, Thailand adapted and implemented HIVQUAL to fit the Thai national HIV guidelines, resources, and health care system [6,7]. A Pediatric HIVQUAL-T was subsequently adapted for use in pediatric HIV patients in Thailand. Their model was first used in five pilot-site hospitals in 2005 and expanded to 10 other hospitals in 2010 and has been a very useful tool for ongoing quality improvement (QI) using hospital data on pediatric HIV care. Through their analyses, several areas needing development (ex. increasing assessment of immunization and dental history) were identified and have been successfully enhanced to provide a positive impact on the quality of patient care [5]. The HIVQUAL software was introduced in Haiti along with an electronic medical record (EMR) system to a few hospitals in 2007 [9,10]. All comprehensive data from HIVQUAL reported so far have been mostly focused on adults or overall adult/pediatric combined data with limited pediatric details (see Appendix 1 for sample report) [9]. There is a need to analyze more pediatric specific data such as adherence to treatment, growth and nutrition assessments, hospitalizations and mortality in order to improve care and outcomes. A dedicated comprehensive pediatric assessment adapted for pediatric HIV care in Haiti, HIVQUAL-H, would be an essential quality improvement measurement tool.

The setting

Saint Damien Hospital, the largest pediatric hospital in Haiti, is located in the town of Tabarre, just outside of the capital of Portau- Prince. The facility has a total of 150 inpatient beds and on-site outpatient clinics, including a clinic dedicated to HIV-infected children. About 800 HIV infected children aged 0 to 15 years are seen yearly in the outpatient HIV clinic [11]. This HIV clinic has had an EMR system (Isanté - Seattle, Washington) since 2009 which incorporates portions of the HIVQUAL-based software (termed HEALTHQUAL since 2013). The latest version includes additional indicators more specific to the adult population such as family planning and uterine cancer screening [10]. Using this EMR and HEALTHQUAL, several abstract presentations of QI projects at St. Damien were presented in 2011 and 2013 and were based on limited HIV care indicators (CD4 count, immunization status, and ART) [12-14]. A few other QI projects were also done at this facility leading up to 2016, but not focused on specific quality of care indicators of pediatric patients (ex. level of adherence among all HIV patients, clinical retention of HIV patients, mental health evaluation in HIV patients, re-evaluation of patients in the emergency room prior to admission, community dispensing of ART). There are no published comprehensive analyses of the specific pediatric HIV care at Saint Damien nor in the country.

The aim of our study was to adapt the HIVQUAL model that was successfully used in Thailand for pediatrics to include countryrelevant pediatric guidelines and indicators to provide a comprehensive overview of the adequacy of Haiti’s pediatric HIV care while providing quality measures that can be used to optimize that care.

Materials and Methods

This is a retrospective chart review of electronic medical records of active and infected pediatric HIV patients less than 15 years of age from 2012 to 2016 from the pediatric HIV outpatient clinic at Saint Damien Hospital in Tabarre, Haiti. HIV-exposed infants and HIV-infected persons age 15 and above were excluded (patients who turned 15 during any year had their data used through that calendar year only). The EMRs of all eligible patients were reviewed for basic demographic information (age and sex), where the rounded age in years was determined by using age on the final day of the year. Using a combination of previous HIVQUAL models [5-8] and Haiti’s national pediatric HIV care guidelines [3] a pediatric HIVQUAL-H model was generated (Table 1) to collect relevant clinical and laboratory indicators. Core indicators include clinical status monitoring (World Health Organization [WHO] HIV staging), CD4 count, viral load monitoring, growth assessment, Pneumocystis jiroveci pneumonia (PJP) prophylaxis, clinical tuberculosis (TB) screening, ART, and ART adherence monitoring. Data were also collected on expanded indicators including immunization history, oral health assessment, HIV disclosure to child, and number of hospitalizations.

Pediatric HIVQUAL-H

Core indicators

Clinical staging was based on the 2007 World Health Organization classification with stage 1-4 [15]. Only the latest staging classification was recorded for each calendar year. Laboratory monitoring indicators recorded were CD4 counts and viral loads obtained per calendar year; both the number of checks and the worst values were recorded from each year. CD4 values were recorded as percentage for patients <6 years and as absolute values for those ≥ 6 years [16-18]. Viral loads were recorded in copies per mL. Growth assessment included the height in centimeters and/or weight in kilograms when measured in a calendar year with only the most recent values recorded. Clinical TB screening was documented based on completion of required yearly screening questions, regardless of screening outcome. Both ART and Pneumocystis jirovecii pneumonia (PJP) prophylaxis were assessed based on relevant medications listed in EMR or not during a given calendar year. In the EMR, adherence is tracked in a special adherence visit note; a standard of 4 completed adherence visit notes per year met the adherence criteria (Table 1).

Expanded indicators

According to the Haitian HIV guidelines, HIV-exposed and asymptomatic HIV positive children should receive the following vaccines by the age of 2 years: Bacille Calmette-Guerin (BCG), oral polio vaccine (OPV) x3, Pentavalent (Diphtheria-Tetanus-acellular Pertussis [DTaP], Haemophilus influenza type B [Hib], Hepatitis B [Hep B]) x3, Rotavirus x2, Measles-Rubella (MR) or Measles-Mumps-Rubella (MMR) x1 [3]. Only patients with documentation for meeting all the parameters above according to their age were counted as up to date on their vaccines. Other documented but not required vaccines were also tallied, especially pneumococcal vaccines, including pneumococcal conjugate vaccine-7 (PCV7) and pneumococcal polysaccharide vaccine-23 (PPSV23).

Oral health assessment was counted as completed for a calendar year if reported in the exam portion of a visit note and any mention of dental discoloration, caries, or abnormal finding was counted as evidence of dental disease. Referral to a dentist was recorded when documented in the plan portion of a visit note for a given calendar year. HIV disclosure status was recorded according to the documentation in visit notes. The target age for disclosure was age 10 based on the policy of the local institution which is consistent with an average age for disclosure as found in reviews of multiple studies across the world (age 8-12 years) [19,20]. Hospitalizations and associated diagnoses were recorded based on documentation within the assessment and plan of the HIV clinic visit notes from the year. Each hospitalization was tallied separately, even if the same child was admitted several times in the same year.

Collected data were aggregated by calendar year per patient for longitudinal comparison. Each patient was assessed for meeting each indicator separately for each eligible calendar year data were available. This study protocol was reviewed and approved by both the University of Nebraska Medical Center (UNMC) Institutional Review Board as well as the Haitian Ministry of Health National Bioethics Committee. Study data were collected and managed using the REDCap electronic data capture tools hosted at UNMC. Service and support was provided by the Research Information Technology Office, which is funded by the UNMC Vice Chancellor for Research.

Statistical Analysis

Descriptive statistics (counts and percentages, means, standard deviations, medians, minimums and maximums) were used to summarize the data. For each Pediatric HIVQUAL-H indicator, the proportion of patients meeting the indicator definition was determined separately for each calendar year. Fisher’s exact test was used to see if there were associations between performance and time period. The Cochran-Armitage trend test was used to test if there was a linear trend in performance during the five-year period. The level of significance is set at P-value <0.05. Statistical analyses were carried out using SAS Version 9.4 (Cary, North Carolina, USA).

Results

Over the five-year period, there were 393 unique patients, with 283 patients in 2012, 291 in 2013, 290 in 2014, 298 in 2015 and 311 in 2016 accounting for 1473 patient-years (PY) analyzed (Table 2). Less than 50% were males (699/1473 PY) with a median age of 8 years and range of 0-15 years (Table 3).

Core indicators

Overall, 96.8% received clinical status monitoring (1426 PY). In 2012 and 2013 most patients (70.3% and 52.9%, respectively) were classified as stage 2 while in 2015 and 2016 (41.9% and 37.2%, respectively) were classified as stage 3 per WHO guidelines (Table 3). Nutritional assessment was only able to be based on body mass index (BMI). Of those with recorded growth parameters, low BMI (<16) ranged from 29-40% by year. There were insufficient data to calculate BMI percentiles. About 6.5% (85/1310 with measured values) had a clinically low CD4 count (<15% in those 1-6 years and <200 in those age 6 and above) during the study period. Overall, 89.8% received yearly CD4 monitoring (1323 PY). Viral load monitoring was only done 50% of the time (730 PY) with frequency increasing in recent years (Table 2). Of those with measured viral loads, 34.9% (255/730) were detectable at least once in any given year (>300 copies/mL being the local cutoff). Viral load was detected in 5.4-17.2% patients yearly except in 2014 when it was detected in 38.3% of patients (Table 3). 94.6% received antiretroviral treatment (1394 PY). Linear trends were noted for viral load, CD4 count, antiviral therapy (all p<0.0001), and clinical staging (p=0.02) (Table 2).

Overall, there was 99.4% PJP prophylaxis (1465 PY), 98.5% TB screening (1452 PY), and 98.7% growth assessments (1454 PY). There was an overall medication adherence monitoring rate of 92.5% (1362 PY). The frequency trend over time showed an overall decreasing percentage of meeting the adherence criterion over the five-year period of the study (p=0.0417, Table 2). There were also some patients not on ART with documented adherence assessments for prophylaxis (76 PY) and not ART, but this was incorporated in the final totals together as it did assess their compliance to the prescribed regimen. We initially aimed to count adherence by the number of adherence notes present; however, on review, several of these templates were completely blank without any actual adherence data recorded (overall around 10%), so we assured at least 4 completed assessments when counting those meeting this criterion.

Expanded indicators

Overall, immunizations were not reliably documented in the EMR and no patient met the status of being “up to date” per the Haiti HIV guidelines (Tables 2 and 4). The most documented vaccines were DTaP and OPV (both >70% each year) and any form of measles vaccine (around 65% each year). BCG was documented in 40-45% of patients. Hep B, Hib, and Pneumococcal vaccines were all documented roughly at or below 33% each year (Table 4).

Though oral health screening rates were high (99% overall) with 44.9% having documented dental disease (661 PY), only 6.1% of those (40/661) had a documented dental referral (Figure 1). Overall, only 30.7% (181/589) of patients aged 10 years and older were disclosed to their diagnosis, with widely-variable yearly rates from 16-45% across the time period (Table 2). The overall hospitalization rate was 4.5% (66 PY, Table 2) accounting for a total of 79 hospitalizations during that time with malnutrition, TB, and HIV complications accounting for the majority of the reasons (67/111, 60.4%) for admission (Figure 2).

Discussion

The proposed HIVQUAL-H model provided a good overview of the quality of pediatric HIV care provided at the St. Damien HIV clinic and highlighted some successes and areas that could use improvement.

Core indicators

The documentation of the WHO clinical staging was well-done likely due to the fact that it is built into the EMR template as a checkbox, which makes it much easier to complete at each visit. CD4 monitoring used to be the mainstay of the monitoring at this clinic, but has trended down in recent years (from 100% in 2012 to 54.6% in 2016) in favor of increased viral load testing, which became more available in 2014- 2016. This also explains why very few had viral loads done in 2012- 2013 (under 50%) and that viral load checks became more frequent by 2016 (88.7%) (Table 2). Prohibitive cost may impede more frequent testing which in some settings allows for closer monitoring of response to therapy. There was excellent documentation of all children receiving PJP prophylaxis, regardless of whether they were on ART or not, which is crucial given this high-risk population.

TB screening rates were high, likely related to the pre-written questions within the EMR note templates. However, those with positive screening questions did not have good documentation of subsequent work-up and treatment, likely because this was done in the inpatient setting and not recorded in the outpatient EMR. We also discovered that during the study period, some patients were placed on empiric daily isoniazid (INH) prophylaxis as per the guidelines of the WHO that came out near the beginning of the study period. The recommendations stated to give at least 36 months of daily INH to HIV patients in TBendemic regions to help prevent the risk of TB infection in these populations [21]. We recorded the number of patients on any length of isoniazid therapy in each given year (Figure 3).

The rate of the active HIV patients on documented ART therapy was very high, above 90% in all years analyzed with significant trend that increased steadily each year (p<0.0001) indicating excellent work on this indicator by staff in recent years. Some patients not on ART were either very recently diagnosed or waiting for a more stable social situation before being started. Though the overall rate of adherence was high (91.0-95.4% by year over 5-year period), this was only looking for four adherence assessments per year. The standard should really be medication adherence assessment at every clinic visit, though this was difficult to discern as not every patient is seen at the same interval, depending on clinical status and history of compliance. Also, given the adherence assessment is listed separately from the visit notes, the documentation was poor, many templates were not filled out and in our analysis only 4 completed forms from total available in a year counted as adherence. In addition, the percentage of patients meeting this criterion decreased each year, suggesting poorer documentation over time. A more thorough evaluation of the adherence documentation processes could lead to improvement in completion rates and increased amount of true assessments at each visit.

Expanded indicators

Immunizations were not reliably recorded and no active patient met the definition of being up to date on vaccines according to the local guidelines. The immunization records were better documented for younger patients, with more adolescent patients often having blank records. Though accurate documentation may be an important issue to address, reliable supplies of routine vaccines may be more of a problem and newer vaccine formulations are only slowly rolling out with Pentavalent (DTaP-Hib-Hep B) vaccine coming in 2012 and Rotavirus not until 2014 [22]. Additional important vaccines in HIV infected population include the pneumococcal vaccines, which are yet to be available routinely and are not yet part of the national guidelines.

Though there was good documentation of oral health assessment, there was a large amount of dental disease described, but not much documentation of dental referrals. No specific dental records were available as the on-site dental clinic where most referrals were sent does not use the EMR, but separate paper charts. Incorporating referrals and completion of these visits would help improve quality of care.

The disclosure rate is very low compared to the hospital goal (100% disclosed by age 10 years) In discussion with local faculty and families, there is still significant stigma surrounding HIV diagnoses in the community. Parents are afraid to tell their children why they take medications as they worry they may inadvertently share this knowledge with friends and neighbors, which could negatively impact their dayto- day lives. There have been improved efforts over the past couple of years at St. Damien for a more step-wise approach to disclosure with families, including the creation of group activities for children and parents as well as dedicated staff including a clinical psychologist to assist in helping and educating families.

The method of recording the number and reasons for hospitalizations was not reliable and the hospital inpatient records are not within the EMR. An integrated hospital/clinic EMR may address this issue and improve tracking of patient outcome. Though we were able to estimate BMI data based on some recorded heights and weights, a more accurate BMI percentile per age was not feasible due to lack of all necessary data (height, weight and age in months) being recorded reliably within the same visit for most of the patients; therefore a BMI <16 was used to loosely indicate malnutrition. A more detailed evaluation and documentation process regarding nutritional status is needed, preferentially to incorporate calculations of WHO Z-scores to allow for the most accurate identification and categorization of at-risk children. This is particularly important as malnutrition was found to be the leading cause of documented hospital admissions in this population during the study period (Figure 2).

Strengths

This study had several strengths, foremost being that it provides the first published data regarding a comprehensive assessment of the quality of pediatric HIV care in Haiti. As it was implemented in the largest pediatric center in the country and five years of data were analyzed, the sample size is robust. As well, the data gathered have clearly identified specific areas within important pediatric HIV care categories to develop in this particular facility. There is now a baseline to design targeted interventions to improve areas of deficiency and allow for repeat assessment.

Limitations

As in any retrospective study, our data were limited to what had been previously recorded in the EMR and therefore, there were several areas that were incomplete or missing. Overall though, given the large pediatric HIV population sampled, the amount of missing elements was more minimal (except in immunizations as discussed above), so we feel the data for the quality of care indicators was still highly accurate.

We only examined the patients currently listed as “active” in their system, which did not allow full assessment across the entire patient population (those who were in the system as inactive due to extended time since last clinic visit, those still with HIV but not on therapy, or those who had been treated during the study period but had since passed away). Therefore, their system’s classification did not allow us to accurately assess mortality rates, as patients that may have been treated in earlier years of the study and died before the time of data collection were not analyzed.

Only the most recent or worst clinical data were recorded from each year where one patient had multiple data values in certain categories. Furthermore, as each patient was assessed separately by calendar year for meeting criteria and these data were later pooled by year, this did not allow for analysis of progress/worsening of individual patients over time.

The current HEALTHQUAL software (Isanté) being used at St Damien is not set to generate reports specific to and in enough detail for pediatrics (see Appendix 1 for sample HEALTHQUAL report and definitions). It would help in the future if the EMR could be adapted to allow for formal reports according to the proposed HIVQUAL-H parameters outlined here allowing for easier and more accurate ongoing surveillance of pediatric data for further improvements.

Future Direction

Immediate next steps should include in-depth analyses of current documentation processes to identify points in the current system that could be quickly improved to address identified deficiencies, such as recording weight and height and age for proper nutritional assessments at each visit. An important next step would be implementing programs with psychosocial support to help in approaching disclosure of HIV status by the age of 10. We should work towards integrating the HIVQUAL-H model into the EMR at St. Damien.

Conclusion

This study attempts to address the gaps in knowledge regarding the quality of care monitoring of pediatric HIV patients in Haiti by applying a tool with core and expanded HIV indicators for assessment customized to local guidelines. This assessment shows performance rates of many pediatric HIV indicators in Haiti are good, whereas some need improvement. Viral load monitoring, oral health care, HIV disclosure, and immunizations are areas that could use focused interventions at this facility to improve standard of care for this population.

We think this tool can be easily adapted for other pediatric HIV programs in developing countries with similar resource constraints. The ability to identify these trends will allow for identification of areas needing improvement and future quality improvement projects, ultimately leading to better care for these patient populations.

Acknowledgements

Special thanks to Judens Bastien, Saint Damien HIV clinic information technology specialist for his help in facilitating data collection and assistance with technical issues during the project.

Funding

This work was supported by The University of Nebraska Medical Center Department of Pediatrics Departmental Grant and Global Health Program.

References

1. U.S. President's Emergency Plan for AIDS Relief (PEPFAR) (2017) FY 2017 Haiti country operational plan (COP), Strategic Direction Summary.
2. U.S. President's Emergency Plan for AIDS Relief (PEPFAR) (2015) FY 2015 Haiti country operational plan (COP).
3. Francois K, Balan JG, Gautier J, Rouzier V, Desinor OY, et al. (2015) Manual of norms for the clinical and therapeutic management of infants and children infected or exposed to HIV.
4. U.S. President's Emergency Plan for AIDS Relief (PEPFAR) (2018) FY 2018 Haiti country operational plan (COP), Strategic Direction Summary.
5. Lolekha R, Chunwimaleung S, Hansudewechakul R, Leawsrisook P, Prasitsuebsai W, et al. (2010) Pediatric HIVQUAL-T: Measuring and improving the quality of pediatric HIV care in Thailand, 2005-2007. Jt Comm J Qual Patient Saf 36: 541-551.
6. Thanprasertsuk S, Supawitkul S, Lolekha R, Ningsanond P, Agins BD, et al. (2012) HIVQUAL-T: monitoring and improving HIV clinical care in Thailand, 2002-08. Int J Qual Health Care 24: 338-347.
7. Ningsanond P (2008) Partnerships for national scale-up of the HIVQUAL-T model for quality improvement in HIV care. XVII International AIDS Conference, Mexico City.
8. Agins BD, Young MT, Ellis WC, Burke GR, Rotunno FF (1995) A statewide program to evaluate the quality of care provided to persons with HIV infection. Jt Comm J Qual Improv 21: 439-456.
9. http://healthqual.org/haiti Demes JAE (2014) Overview of a continuous quality improvement program in Haiti. Quality improvement Health Haiti.
10. http://www.andreabocellifoundation.org/projects/hiv-program-st-damien-pediatric-hospital/
11. Excellent ML, Dorgelus D, Gautier J, Severe D (2013) Enrollment on HAART at Hôpital Saint Damien /Nos Petits Frères et Soeurs (HSD/NPFS), Haiti.
12. Augustin MC, Phadaël S, Goldman MAB, Dubuisson PD, Metellus V, et al. (2013) Improving pediatric immunization at Hôpital Saint Damien /Nos Petits Frères et Soeurs (HSD/NPFS).
13. Pierre E, Romulus CV, Pierre FC, Pierre J, Cenet C, et al. (2013) Improving CD4 monitoring at Hôpital Saint Damien/Nos Petits Frères et SÅ?urs (HSD/ NPFS), Haiti.
14. World Health Organization (2007) WHO case definitions of HIV for surveillance and revised clinical staging and immunological classification of HIV-related disease in adults and children.
15. Denny T, Yogev R, Gelman R, Skuza C, Oleske J, et al. (1992) Lymphocyte subsets in healthy children during the first 5 years of life. JAMA 267: 1484-1488.
16. Shearer WT, Rosenblatt HM, Gelman RS, Oyomopito R, Plaeger S, et al. (2003) Lymphocyte subsets in healthy children from birth through 18 years of age: The Pediatric AIDS clinical trials group P1009 study. J Allergy Clin Immunol 112: 973-980.
17. US Department of Health and Human Services (2018) Guidelines for the use of antiretroviral agents in pediatric HIV infection: Clinical and Laboratory Monitoring of Pediatric HIV Infection. AIDS Info.
18. Wiener L, Mellins CA, Marhefka S, Battles HB (2007) Disclosure of an HIV diagnosis to children: History, current research, and future directions. J Dev Behav Pediatr 28: 155-166.
19. Britto C, Mehta K, Thomas R, Shet A (2016) Prevalence and correlates of HIV disclosure among children and adolescents in low- and middle-income countries: A systematic review. J Dev Behav Pediatr 37: 496-505.
20. World Health Organization (2011) Intensified tuberculosis case-finding and isoniazid preventive therapy for people living with HIV in resource-constrained.
21. Tohme RA, Francois J, Cavallaro KF, Paluku G, Yalcouye I, et al. (2017) Expansion of vaccination services and strengthening vaccine-preventable diseases surveillance in Haiti, 2010-2016. Am J Trop Med Hyg 97: 28-36.